INTRODUCTION

There is an ongoing interest in analyses of various contaminants in biological material ranging from plasma, serum, hair, nails to soft and hard tissues. These analyses not only have certain value for basic research, but they can also serve as a biomarker of human exposure to environmental pollution or occupational risks [1]. Particular interest is given to their potential prognostic or diagnostic value.

So called “heavy metals” are one of the most studied groups of environmental contaminants with number of elements evidenced to display toxic effects on reproduction [2]. The contractual group includes non-essential toxic metals such cadmium (Cd) and lead (Pb) which can induce spontaneous abortion and have been implicated in fertility issues, endometriosis and reproductive cancer [3]. The other, e.g., zinc (Zn), copper (Cu), manganese (Mn) and chromium (Cr) are pivotal for reproduction but can lead to harmful outcomes both, at deficiency and at increased levels [4, 5]. The additional attention is also paid to metalloids such as arsenic (As) and light metals such as aluminum (Al). The latter is ubiquitous in different foodstuffs and widely used in medicines [6]. It can exert metalloestrogenic action as observed in vitro, and lead to endocrine alterations evidenced in vivo [7, 8].

Continuous development of analytical procedures enables analyses of various biological material and determination of wide range of chemical elements. Apart from fluid samples, recently several studies directly assessing metal concentrations in tissues collected from women reproductive trait have been performed [9–13].

The studies concerning the role of toxic metals in female human reproductive organs has been performed in our department since 2015. The aim of the present study was to use the results accumulated in our database on concentrations of selected metals in human endometrium (differentiated by histological conditions) and chorionic tissue in evaluation whether they could have any prognostic value [12, 13]. Therefore, statistical calculations based on discriminant and canonical analyses were performed.

MATERIAL AND METHODS

This paperwork is an additional statistical analysis of data largely already published. Thus, the calculations were performed retrospectively on previously partly published data [12, 13]. It included concentrations of lead (Pb), cadmium (Cd), nickel (Ni), manganese (Mn), copper (Cu), chromium (Cr), Zinc (Zn) and aluminum (Al) in human endometrium [11, 12]. The following groups were divided: normal endometrium (HP1, n = 28), endometrial hyperplasia (HP2, n = 15), endometrial cancer (HP3, n = 5), miscarriages (HP4, n = 19) and endometrial polyps (HP5, n = 25).

All patients have undergone curettage of the uterine cavity due to abnormal uterine bleeding, miscarriage or other suspicious changes such as endometrial hyperplasia or endometrial polyp. After that histopathological examination was conducted.

Microwave induced plasma atomic emission spectrometry was used to assess metals concentrations in tissue material and the results of obtained concentrations were converted to µg/kg.

The calculations were performed using discriminant and canonical analyses by Statistica 10.0 software (StatSoft, U.S.A.). Assumptions for the model concept were Tolerance: 0,01, Probability enter = 1, Probability removed = 0, Steps = 8. Canonical analyses revealed four discriminant functions. p < 0.05 was considered as statistically significant.

RESULTS

`The studies revealed different concentrations of metals in tissues with different histopathological diagnoses (Tab. 1).

When all metals were introduced to discriminant function analysis, Zn, Cd and Cr were found not to meet the model’s concept and had the smallest contribution to the discrimination between groups of diagnosis. Therefore, these elements were excluded from further analyses. The discrimination of histopathological diagnosis by five metals that were included in the model (Cu, Pb, Ni, Mn and Al) was highly significant (Wilks’ Lambda: 0,09580, p < 0.05). Pb and Ni have the smallest partial Wilks’ Lambda of 0.37 and 0.53, respectively, thus the highest discriminant power for this model (Tab. 2). The high tolerance values were observed for Cu and Mn, indicating that 70 and 60% of information brought by these variables, respectively, was not provided by the other variables included in the model (Tab. 2).

Then canonical analysis was performed to give four discriminant functions (Roots). Three of them were statistically significant (p < 0.05) (Tab. 3).

The first function discriminates most of all endometrial hyperplasia from the miscarriage group and endometrium without pathological changes and has the highest discriminant power (2.81). Furthermore, Pb and Cu play the most crucial role in this discrimination. The second function refers principally to endometrial cancers mainly based on Mn and Cu concentrations. The third one (with the lowest discriminant power and with the lowest statistical significance) distinguishes the miscarriage group and endometrium without pathological changes (Tab. 4 and 5).

Figure 1. The configuration of points representing histopathological diagnosis in the system specified by the variable (axes) of discrimination; HP1 — endometrium without histopathological changes; HP2 — endometrial hyperplasia, HP3 — endometrial cancer; HP4 — miscarriages; HP5 — endometrial polyp

Taking into account the whole studied group, different concentrations of metals have the cardinal impact on discrimination between histopathological diagnosis.

Pb had the greatest impact on classifying a tissue sample to histopathological diagnosis in case of endometrial hyperplasia, endometrium from miscarriages and endometrial polyps. In case of endometrial cancer such value was observed for Al (Tab. 6).

As indicated, 67% of tissue samples were correctly classified to their histopathological diagnosis groups, based on the discriminant analysis.

DISCUSSION

The present study has revealed using discriminant analysis that based on metals concentrations in tissue, it is possible to classify it sample to the histopathological diagnosis (67% corrected classified samples). It is suggested that there may be a different scheme of metals concentrations in order to reach a different histopathological diagnosis. These might be primary or secondary changes. It is unknown whether concentration of metals in endometrium is primary findings that may cause tissue alteration or secondary accumulation due to histopathological lesions.

The first canonical function that discriminated mainly endometrial hyperplasia samples from endometrium without pathological changes and miscarriages group was based on Pb and Cu concentrations. Pb is considered as metaloestrogen — metal that can join estrogens receptors and stimulate it. It may have estrogenic properties, which are the main cause of endometrial hyperplasia [14].

There is evidence that also Cu may play a role in endometrial hyperplasia [15]. Chianeh and Rao suggest it may occur through vascular endothelial growth factor (VEGF) and endothelial nitric oxide synthase (eNOS) [16, 17]. It is pointed out that angiogenesis requires endothelial growth and differentiation.

Second canonical function referred to endometrial cancer samples is based on Mn and Cu. There are considerable findings revealing that oxygen stress is one of the most important causes of cancerogenic process [18, 19]. Mn and Cu are cofactors of stress oxygen dismutase (SOD) enzyme which takes part in defense mechanism, and this might be a reason that these two are so important in that discrimination. More studies provide extensive evidence that SOD activity is decreased in malignant cancers [20]. On the other hand, some authors suggest it depends on a type, stage and grade of a cancer. It is also assumed that SOD level may fluctuate during cancerogenesis [21, 22].

Taking the chart into consideration, HP5 and HP1 are almost in quite the same place, so these two groups are the most difficult to discriminate from each other based on metals concentrations. Assuming the chorion is created from endometrium, and it is new young tissue (till first 12 weeks of pregnancy), it may be concluded that it would differ from normal endometrium in the smallest extant. Other 3 groups are spread on the chart which indicates that the concentrations of toxic metals may be important.

All in all, it seems that the crucial metal in whole discrimination process is Pb while in case of endometrial cancer — Al. Pb is well known toxic and cancerogenic metal. It may play a decisive role also in endometrial cancerogenesis. Al, in turn, is considered as metalloestrogen and it seems it can stimulate for example, oestrogen receptors of MCF7 human breast cancer cells through the ligand binding or oestrogen-regulated reporter gene expression [23]. Additionally, the studies revealed that bladder and lung cancer is associated with aluminum smelting [24, 25].

In literature, there are evidence for specific, characteristic toxic metals concentrations in different types of diseases. Yu Yuan [26] discovered that five metals: titanium, arsenic, selenium, aluminum, and barium are significantly present in the blood of patients with coronary heart disease. Xiji Huang et al. [27] presented that asthma prevalence in adults is associated with urinary chromium, chromium, selenium, molybdenum, cadmium, and uranium concentration. Moreover, urinary lead, barium, iron, zinc, nickel, manganese and rubidium were negatively associated with asthma and interestingly when authors compare urinary metals in different subgroups, the associations of above 13 metals with asthma prevalence were nearly the same [27].

In our study, discriminant model enabled to classify 67 % of our samples to their correct histopathological diagnosis. It is possible that metals concentrations in blood, urine or tissue may serve us as a prognostic, additional examination in a diagnostic process.

Additional research with larger amounts of tissue samples is required to support our findings. Another limitation of the study is fact, that not all known toxic metals have been tested in the tissues, such as mercury. There may be more trace elements related to the metal’s schemes.

CONCLUSIONS

Acknowledgments

We would like to express sincere thanks to Professor Piotr Rzymski, PhD Barbara Poniedziałek, Professor Przemysław Niedzielski for the identification and assessing of metal concentrations in tissue samples.

Conflict of interest

All authors declare no conflict of interest.